Machine Learning

I am interested in machine learning and I think that it is going to change the world around us before we know it. At my current job, we are already seeing people replaced by automation. Machines are outperforming people in so many different ways. Intelligent systems never forget anything and they are more accurate than humans. The crazy scary fact of the matter is that they are now outperforming our cognitive abilities. I see many benefits for education and performance improvement, but I am curious about other applications for machine learning. I stumbled across this article with the title, ‘Machine Learning Will Change Medicine’ and wanted to share some of their thoughts.

The authors described medical imaging as simple pattern recognition and claimed that we are right on the verge of letting machines take care of this. They go on to describe how machine learning will eventually lead to diagnoses that are more accurate and possibly help us reduce costs. Doctors overlook things all of the time, because well, they are humans. The authors claim that many so-called rare diseases are usually just overlooked - doctors don’t always have the cognitive capacity to remember all of these rare cases that they learned about in medical school. I thought that this was quite interesting and made a lot of sense.

It turns out that there are many barriers to using machine learning in the medical field. The main issue is that they do not currently have enough medical datasets in their cloud to train the systems at this time. Nevertheless, they are working on it and it is coming soon! I thought that it was funny that one of the issues discussed was that there would be no one to hold responsible for mess-ups caused by the machines. I quickly realized that this is all going to change the way we look at things in our world. How would you feel if a machine gave you an improper diagnosis? Would it be a different feeling than if a human misdiagnosed you? I think that this is all so very cool, but it almost seems too futuristic at this point. Can you all think of other facets in life where machine learning could possibly be used to improve our society?
Check out the article if you have time - it’s a short read!   http://jnm.snmjournals.org/content/58/3/357.short

Comments

  1. UnknownJuly 8, 2017 at 9:30 PM

    This comment has been removed by the author.

    ReplyDelete
  2. UnknownJuly 8, 2017 at 9:32 PM

    One thing I’ve seen as progress in the automation field relating to something I do is content and image correlation. Reading content and finding a relevant graphic or photo is a time-consuming part of my job. I’ll spend hours scouring the web for pictures or building graphics that match content. Computes are getting better at making associations and identify objects. So, looking at pictures and identify age or race is something a computer can do now. Currently we use meta data but that all has to be manually entered into the system. Computers are also becoming more efficient at reading materials and interpreting meaning from the text. Automation can fact check and credit source in the blink of an eye. Automation makes everyone’s job a whole lot easier. I can foresee a time when automation scans content on web pages and correctly identifies or generates digital assets that aim to improve user experiences.

    Check out what google is doing with Quick, Draw here: https://quickdraw.withgoogle.com/ They are trying to build the world’s largest data set of doodles.

    And this cool Image-to-Image Demo here: https://affinelayer.com/pixsrv/index.html
    Where patterns and drawings generate pictures based on the composition of the input.

    ReplyDelete

Post a Comment

Popular posts from this blog

Reflection - Web 2.0 Tools

The benefits of constructivism, discovery learning, connectivism, and social learning are absolutely phenomenal. I have quickly noticed that these Web 2.0 tools facilitate learning through all of these cognitive theories. I think that this class has highlighted the importance of collaboration and the benefits of cognitive conflicts (diversity of opinions) that come with social learning. It has been interesting to see how knowledge can be created, synthesized, and shared more effectively and efficiently using these Web 2.0 tools. I also feel like these tools are great for role-playing activities and can be used for online learning games. I believe that learning outside of the traditional ‘formal’ environment can breed more authentic experiences and that the students will have access to a more diverse audience when they use these tools. Over the last four weeks I have noticed that learners tend to take more control of their learning when using these web 2.0 tools and that there are ...
Read more

Wiki Use for Virtual Microscopy and Pathology Reports

I found this cool article that reviewed the use of wikis at the New York College of Podiatric Medicine. They have a pathology program there where they created several wikis for educational purposes. They use virtual microscopy where students scan microbial slides by various sections and then other learners can manipulate the images by zooming in, underlining things, adding arrows, labeling attributes, etc. Students group images by topic and are even able to edit other student’s slides. This is all preparing them to work as pathologists and I thought that it was a unique way to develop these professional skills. Students also participate in discussions on the wiki and share new ideas with each other.  The college discovered that this method of instruction produced an extremely detailed analysis of the subject matter. It was also beneficial because the students could access this archive at any time for knowledge management. The instructors reviewed student access and tracked...
Read more

My Journey in Curation – How to Catch a Redfish

Image
When beginning this project, I used Pinterest as a repository for the collection of resources. This made it easy for me collect resources on a ‘board’ for easy access and analysis. In total, I collected nine different sources of information - some content was better than others and there was also a lot of overlap between sources. I mined these nine sources for information that I thought would be beneficial for novice anglers going after Redfish for the first time. First, I wanted to begin the curation process with an instructional Youtube video that I personally created. The video teaches learners how to adequately identify a Redfish. I figured that this is the most important step in learning how to catch a Redfish, so this seemed like a natural starting point. In this video I explain: ·          How to identify the copper-bronze or red body color on the top half of the fish ·          How to ident...
Read more